339409 - Planet RSS feeds need to specify an encoding and contains illegal characters

Status: RESOLVED FIXED

(Websites :: planet.mozilla.org, defect)

Description

[Phil Ringnalda (:philor)]

planet.mozilla.org sends its RSS feeds as text/xml, with no charset param, and with no encoding specified in the XML decl. You couldn't get a better recipe for mangled encoding: a consumer who believes in RFC 3023 will treat it as US-ASCII, and explode, a consumer who believes that ISO-8859-1 is the default for HTTP will treat it as that and mangle it (which looks like the current state of Firefox's feed preview), and a consumer who believes that UTF-8 is the default for all XML, not just application/xml, will treat it as that.

I like belts and suspenders, so I'd add both a ; charset=utf-8 to the content-type header and a <?xml version="1.0" encoding="utf-8"?>, but either one ought to be enough, if it's also served as application/xml.

Comment 1

[Robert Sayre]

The feed also contains illegal characters. Suggest upgrading to the newest version of planet.

Comment 2

[tor]

I've updated to the current version of planet, which tags the html and atom feed with utf-8.

Comment 3

[Robert Sayre]

http://feedvalidator.org/check.cgi?url=http://planet.mozilla.org/rss20.xml

encoding looks fine to me, modulo the bogus RFC3023 warnings the validator likes to send.

Comment 4

[Daniel C]

$ curl --head http://planet.mozilla.org/rss20.xml;
HTTP/1.1 200 OK
Date: Wed, 07 Jun 2006 13:08:44 GMT
Server: Apache/1.3.27 (Unix)  (Red-Hat/Linux) mod_ssl/2.8.12 OpenSSL/0.9.6b DAV/1.0.3 mod_perl/1.26
Last-Modified: Wed, 07 Jun 2006 13:06:23 GMT
ETag: "77d43-28241-4486cf4f"
Accept-Ranges: bytes
Content-Length: 164417
Connection: close
Content-Type: text/xml

Can this be changed to be sent as application/rss+xml or application/xml?